Heidelberg 2022 – wissenschaftliches Programm

Bereiche | Tage | Auswahl | Suche | Aktualisierungen | Downloads | Hilfe

AKPIK: Arbeitskreis Physik, moderne Informationstechnologie und Künstliche Intelligenz

AKPIK 2: Data Analytics & Machine Learning

AKPIK 2.6: Vortrag

Mittwoch, 23. März 2022, 17:30–17:45, AKPIK-H13

Simulation of High-Granularity Calorimeter Showers for the ILD Using Normalizing Flows — •Imahn Shekhzadeh — Universität Hamburg, Hamburg, Deutschland

The large computational cost of Monte Carlo simulations together with recent advances in deep learning motivate using deep generative models to speed up simulations. This talk explores the use of normalizing flows (NFs) for high-granularity calorimeter simulations, such as the ones planned for the International Large Detector (ILD). We show that NFs are able to generate high-fidelity showers of simulated photons in the electromagnetic calorimeter of the ILD. Strictly monotonic rational quadratic spline flows are used to enhance the fidelity in comparison to the generally used affine-linear transformations. Finally, we compare the generative performance of the NFs to other state-of-the-art generative network architectures